Machine Learning and Network Medicine approaches for Drug Repositioning for COVID-19. Santos et al.
Description
Supplemental Items used in "Machine Learning and Network Medicine approaches for Drug Repositioning for COVID-19" by Suzana de Siqueira Santos, Mateo Torres, Diego Galeano, María del Mar Sánchez, Luca Cernuzzi, and Alberto Paccanaro. The items include input datasets, and predictions by our matrix decomposition model and network medicine approach. Predictions by our matrix decomposition model are shown in Supplementary File 1. Two out of three interactomes used by our network medicine approach were mapped to SwissProt proteins and shown in Supplementary Files 2 (Luck et al., 2020, DOI: 10.1038/s41586-020-2188-x), and 3 (Cheng et al., 2018, DOI: 10.1038/s41467-018-05116-5). For these two interactomes, we used drug-target associations in Supplementary File 4, and 336 host proteins (UniProt accession numbers) in Supplementary File 5. For the Gysi el al. interactome (Gysi et al., 2021, DOI: 10.1073/pnas.2025581118), we used Entrez IDs of the 336 host proteins shown in Supplementary File 6. Supplementary File 7 shows the ATC categories of DrugBank drugs. DrugBank entries with effect against SARS-CoV-2 according to CMAP, in vitro, and clinical trials evidence are shown in Supplementary Files 8, 9, and 10, respectively. Predictions by our network medicine approach are shown in Supplementary File 11. The weights assigned to the host proteins by the kernel-based methods are shown in Supplementary File 12.